Apertus 1.5 70B
- Family
- Apertus
- Released
- 2026-07-24
- Context
- 262k
- Parameters
- 70B
- Architecture
- Decoder Only
- Specialization
- general
- Openness
- Open source
- License
- Apache 2.0OSI-approvedCommercial use: permitted
- Weights
- Available
- Code
- Unknown
- Training
- Multi-stage
Cheapest of 1 route · Featherless · cache read $0.175
About
Apertus 1.5 70B (HF swiss-ai/Apertus-v1.5-70B) is the Swiss AI Initiative's (ETH Zurich, EPFL, CSCS) fully open 70B dense decoder-only model, released 24 July 2026. It was built by continuing Apertus 1.0 pretraining on another 2T tokens of text and multimodal data, then post-training with SFT, RLVR and DPO. It takes text, image and audio (experimental) input and outputs text, with a 262,144-token context, an optional thinking mode and tool calling. Weights, training data and recipes are public under Apache 2.0; downloading requires an auto-approved Hugging Face click-through (Apache 2.0 + Acceptable Use Policy + contact sharing).
Provider price ladder
Compare API pricing across 1 providers for input and output tokens, batch, and cached reads when available.
| Provider | Input / 1M | Output / 1M | Cache | Route |
|---|---|---|---|---|
| Featherless | $0.875 | $2.08 | read $0.175 | Serverless |
Capabilities
Benchmark peer barsfor RAG
No task-mapped benchmark peers are available for this model yet.
Migration checks
No linked migration route is available for this model yet.
API versions
swiss-ai/Apertus-v1.5-70BApertus 1.5 70Bswiss-ai/apertus-v1.5-70b (Public AI)swiss-ai/apertus-v1.5-70b-thinking (Public AI)Cheapest of 1 route · Featherless · cache read $0.175